Background: Repetition of self-harm is common and is strongly associated with suicide. Despite this, there is limited research on high-volume repetition. Aim: To investigate individuals with high-volume repeat self-harm attendances to the emergency department (ED), including their patterns of attendance and mortality. Method: Data from the Multicentre Study of Self-Harm in England were used. High-volume repetition was defined as ⩾15 attendances within 4 years. An attendance timeline was constructed for each high-volume repeater (HVR) and the different patterns of attendance were explored using an executive sorting task and hierarchical cluster analysis. Results: A small proportion of self-harm patients are HVRs (0.6%) but they account for a large percentage of self-harm attendances (10%). In this study, the new methodological approach resulted in three types of attendance patterns. All of the HVRs had clusters of attendance and a greater proportion died from external causes compared with non-HVRs. Conclusion: The approach used in this study offers a new method for investigating this problem that could have both clinical and research benefits. The need for early intervention is highlighted by the large number of self-harm episodes per patient, the clustered nature of attendances, and the higher prevalence of death from external causes.